Keith Godfrey: Modeling visual system development - a forecast for change?
Talk by Keith Godfrey, University of Cambridge, given to the Redwood Center for Theoretical Neuroscience at UC Berkeley. Presented on September 2, 2009.
Producer Redwood Center for Theoretical NeuroscienceAudio/Visual sound, color
Neural development is a complex process that results from the interaction of many underlying physiological mechanisms. Over the past few decades, computational modellers have produced many models which describe how one or a few of these mechanisms might be responsible for experimentally observed patterns of development. Most of these models, while simple and analytically tractable, have had limited success in providing insight into the biological mechanisms. Climate is another complex process that also results from the interaction of many underlying mechanisms. Jokes about weather forecasters not withstanding, climate researchers are producing increasingly accurate and successful models. Part of this success is because these models represent a wide range of mechanisms that contribute to climate.
In this talk I describe and discuss a model of visual system development which takes the climate modelling approach. The model addresses development of the retinocollicular pathway, when axons first innervate the colliculus, and is based on approximations of many mechanisms, including retinal activity, molecular guidance, trophic factor release, spiking neurons, and the growth and retraction of individual axons and synapses. The core assumptions of the model are that the probabilities of axonal branching and synaptic growth are highest where the combined influences of chemoaffinity and trophic factor cues are highest, and that activity-dependent release of trophic factors act to stabilize synapses. Model axons reproduce morphologically accurate patterns of growth.
Results from the model suggest that spike-timing dependent plasticity and lateral connectivity among collicular neurons are not necessary for retinotopic organization or reﬁnement; that axon growth cones do not require gradient detection during retinotopic positioning or refinement; and that retinotopic development is a two-stage process, mediated first by molecular guidance and then by neural activity. Although complex in detail, the model is insensitive to variations in how the component mechanisms are implemented. These mechanisms are common throughout neural development and the findings of the model, and future models exhibiting similar detail, can shed new insights into neural development throughout the brain.